Mycobacterium tuberculosis is remarkable in its ability to infect the human host and remain quiescent for many years only to reactivate when host defenses are suppressed. One-third of the global population is latently infected with tuberculosis, yet this clinically inactive state when bacilli are often non- cultivatable is poorly understood and inadequately characterized. Appropriate animal models that more accurately mimic human diseases are needed to test vaccine candidates, and to understand the complex relationship between mycobacteria and host. The currently available animal models in mice and guinea pigs are characterized by multibacillary disease and are distinguished by the host response. Mice mount a poor delayed type hypersensitivity response and develop chronic lung disease, eventually succumbing to a progressive granulomatous pulmonary disease with high bacillary load. Guinea pigs also have a multibacillary disease, but are exquisitely susceptible with rapid hematogenous dissemination and a strong DTH response that results in rapid lung inflammation, destruction and death. Paucibacillary latent disease can only be induced with the administration of antibiotics. In contrast, rabbits are relatively resistant to aerosol infection with M. tuberculosis and mount a granulomatous response that effectively contains the bacilli. Over the course of 6-l2 months, all culturable bacilli disappear. In addition, the histopathologic response is remarkably similar to that of humans pointing to the rabbit model of tuberculosis as a promising avenue by which to study stage-specific changes in both host and bacilli. In this application, we will aerosol infect rabbits and allow the granulomatous lung lesions to regress to latency. With the use of immunosuppressive agents such as corticosteroids, iNOS inhibitors, and specific anti-cytokine antibody, we will reactivate infection. In parallel, we will refine the in vitro granuloma assay using rabbit white blood cells to have an in vitro model with which to correlate and compare our in vivo results. Harvesting serum and rabbit tissue at various stages of infection, we will characterize the stage-specific host humoral and cell-mediated immune responses. Understanding the antibody expression profile during latent disease may lead to important diagnostics in a disease hampered by diagnostics with low sensitivity and specificity. In addition, we will analyze the transcriptional bacterial response to various disease stages using microarrays and RT-PCR with molecular beacons. Finally, we will use a transposon mutant library to identify clones that are impaired in specific stages of infection. Appropriate animal models are critical to the successful development of tuberculosis vaccines, new drugs and better diagnostic tests for tuberculosis.